Federated learning for improving matching efficiency

Patent metadata

Field	Value
Publication number	US-12468985-B2
Application number	US-202117461979-A
Country	US
Kind code	B2
Filing date	Aug 30, 2021
Priority date	Oct 16, 2020
Publication date	Nov 11, 2025
Grant date	Nov 11, 2025

Abstract

A method includes: sending, by one or more computers, in response to the number of data providers for federated learning being greater than a first threshold, a data field required for the federated learning to a coordinator, the coordinator comprising a computer; receiving, by one or more computers, from the coordinator, information about one or more data providers comprising the required data field, for determining the data providers comprising the required data field as the remaining data providers, wherein the coordinator stores a data field of each data provider; and performing, by one or more computers, federated learning-based modeling with each of the remaining data providers.

First claim

The invention claimed is:

1 . A computer-implemented method comprising:
sending, by one or more computers, in response to a number of data providers for federated learning being greater than a first threshold, a data field required for the federated learning to a coordinator, the coordinator comprising a computer; receiving, by the one or more computers, from the coordinator, information about one or more data providers comprising the required data field, for determining the data providers comprising the required data field as remaining data providers, wherein the coordinator stores a data field of each data provider; acquiring, by the one or more computers, in response to a number of the remaining data providers being greater than the first threshold, a first public key from the coordinator to encrypt a local user ID, wherein the local user ID is a user ID of a data user; acquiring a user ID, by the one or more computers, which has been encrypted with a second public key, of each of the remaining data providers, wherein the second public key for the remaining data provider is acquired from the coordinator; acquiring, by the one or more computers, an overlap rate by comparing the local user ID with the acquired user ID of each of the remaining data providers; and screening, by the one or more computers, to obtain one or more data providers with an overlap rate being greater than a second threshold, to use the one or more data providers obtained by screening as new remaining data providers; and performing, by the one or more computers, federated learning-based modeling with each of the remaining data providers.

2 . The method of claim 1 , further comprising:
acquiring, by the one or more computers, in response to the number of the remaining data providers being greater than the first threshold, a statistical indicator of respective data fields of the remaining data providers; and performing, by the one or more computers, screening on the remaining data providers based on the statistical indicator, to use one or more data providers obtained by screening as new remaining data providers.

3 . The method of claim 2 , wherein the statistical indicator comprises one or more of the following: a mean value, a maximum value, and a minimum value.

4 . The method of claim 1 , wherein performing federated learning-based modeling with each of the remaining data providers comprises:
screening, by the one or more computers, feature data that is common with each of the remaining data providers by invoking a feature screening model; and performing, by the one or more computers, federated learning-based modeling with each of the remaining data providers based on the feature data.

5 . The method of claim 1 , wherein the required data field comprises a time interval.

6 . An electronic device, comprising:
a memory storing one or more programs configured to be executed by one or more processors, the one or more programs including instructions for causing the electronic device to perform operations comprising: sending, in response to a number of data providers for federated learning being greater than a first threshold, a data field required for the federated learning to a coordinator; receiving, from the coordinator, information about one or more data providers comprising the required data field, for determining the data providers comprising the required data field as remaining data providers, wherein the coordinator stores a data field of each data provider; acquiring, in response to a number of the remaining data providers being greater than the first threshold, a first public key from the coordinator to encrypt a local user ID, wherein the local user ID is a user ID of a data user; acquiring a user ID, which has been encrypted with a second public key, of each of the remaining data providers, wherein the second public key for the remaining data provider is acquired from the coordinator; acquiring an overlap rate by comparing the local user ID with the acquired user ID of each of the remaining data providers; and screening to obtain one or more data providers with an overlap rate being greater than a second threshold, to use the one or more data providers obtained by screening as new remaining data providers; and performing federated learning-based modeling with each of the remaining data providers.

7 . The electronic device of claim 6 , the one or more programs further including instructions for causing the electronic device to perform operations comprising:
acquiring, in response to the number of the remaining data providers being greater than the first threshold, a statistical indicator of respective data fields of the remaining data providers; and performing screening on the remaining data providers based on the statistical indicator, to use one or more data providers obtained by screening as new remaining data providers.

8 . The electronic device of claim 7 , wherein the statistical indicator comprises one or more of the following: a mean value, a maximum value, and a minimum value.

9 . The electronic device of claim 6 , wherein performing federated learning-based modeling with each of the remaining data providers comprises:
screening, feature data that is common with each of the remaining data providers by invoking a feature screening model; and performing federated learning-based modeling with each of one or more remaining data providers based on the feature data.

10 . A non-transitory computer-readable storage medium that stores one or more programs comprising instructions that, when executed by one or more processors of an electronic device, cause the electronic device to implement operations comprising:
sending, in response to a number of data providers for federated learning being greater than a first threshold, a data field required for the federated learning to a coordinator, the coordinator comprising a computer; receiving, from the coordinator, information about one or more data providers comprising the required data field, for determining the data providers corresponding to information comprising the required data field as remaining data providers, wherein the coordinator stores a data field of each data provider; acquiring, in response to a number of the remaining data providers being greater than the first threshold, a first public key from the coordinator to encrypt a local user ID, wherein the local user ID is a user ID of a data user; acquiring a user ID, which has been encrypted with a second public key, of each of the remaining data providers, wherein the second public key for the remaining data provider is acquired from the coordinator; acquiring an overlap rate by comparing the local user ID with the acquired user ID of each of the remaining data providers; and screening to obtain one or more data providers with an overlap rate being greater than a second threshold, to use the one or more data providers obtained by screening as new remaining data providers, and performing federated learning-based modeling with each of the remaining data providers.

11 . The non-transitory computer-readable storage medium of claim 10 , the operations further comprising:
acquiring, in response to the number of the remaining data providers being greater than the first threshold, a statistical indicator of respective data fields of the remaining data providers; and performing screening on the remaining data providers based on the statistical indicator, to use one or more data providers obtained by screening as new remaining data providers.

12 . The non-transitory computer-readable storage medium of claim 11 , wherein the statistical in

Assignees

• Beijing Baidu Netcom Sci & Tech Co Ltd

Inventors

• Chen Lianghui
• Fu Yan
• Zhou Yangjie
• Fang Jun

Classifications

• G06Q40/03 Primary

  Credit; Loans; Processing thereof · CPC title

• H04L9/30

  Public key, i.e. encryption algorithm being computationally infeasible to invert or user's encryption keys not requiring secrecy · CPC title

• G06F16/27

  Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor · CPC title

• G06F16/2365

  Ensuring data consistency and integrity · CPC title

• G06F21/602

  Providing cryptographic facilities or services · CPC title